Simultaneous geomagnetic measurements on an ice island surface and 1000 feet below

Columbia University in the City of New York

LAMONT GEOLOGICAL OBSERVATORY
PALISADES. NEW YORK

SIMULTANEOUS

GEOMAGNETIC MEASUREMENTS
ON AN ICE ISLAND SURFACE
AND 1000 FEET BELOW

by

J. R. Heirtzler

Technical Report No. 2
CU'2'63 Geology
Contract Nonr-266 [82]

May 1963

LAMONT GEOLOGICAL OBSERVATORY
COLUMBIA UNIVERSITY
PALISADES, NEW YORK

S IIvIULTANE OUS GEOMAGNETIC MEASUREMENTS ON AN ICE ISLAND

SURFACE AND 1000 FEET BELOW

by

J* R. Heirtzler

Technical Report No* 2
CU-2-63 - Geology
Contract Nonr-266(82)

May 1963

Digitized by the Internet Archive
in 2020 with funding from
Columbia University Libraries

https://archive.org/details/simultaneousgeomOOheir

CONTENTS

Page

I* INTRODUCTION 1

II. FIELD OPERATION 2

III. GENERAL DESCRIPTION OF RESULTS 4

IV. TIME VARIATIONS 6

V. SPATIAL VARIATIONS 9

VI. CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK 11

APPENDIX 13

ACKNOWLEDGMENTS 15

REFERENCES l6

ABSTRACT

For a few weeks in the fall of 19&2 the total geomag¬
netic field intensity was measured simultaneously on an ice
island surface and approximately 1000 feet below* The mag¬
netic gradient as indicated by the difference between the two
readings varied as the station passed over geologic bodies.

A statistical analysis of the time variations during two time
intervals revealed an attenuation and phase shift of the lower
head reading with respect to the surface head reading. The
analysis was made between 70 and ij.00 seconds period. There
are indications of an anomalous attenuation at the lower
period end of this band although the experiment was not such
that accurate determinations could be made.

ii

I. INTRODUCTION

In the spring of 1962 the U. S. Naval Ordnance Labora¬
tory, White Oak, entered discussions with Lamont*s Arctic Geo¬
physics and Geomagnetism Departments concerning the feasibility
of underwater measurements of the time variations of the geo¬
magnetic field intensity from an ice island. It was jointly
decided that the easiest and most orderly approach to such an
investigation would be the measurement of the total geomagnetic
field intensity with a proton precession magnetometer simul¬
taneously on the ice island surface and at a depth of 1000 feet
below the surface*

It was clearly recognized that the sensitivity and fre¬
quency response limitations of this type instrument would allow
only very gross attenuations and phase shifts to be detected
and such gross effects would probably not be found* Never-the-
less it was felt that this first step should be taken before
more refined studies of the attenuation by sea water were
undertaken. This work should lay to rest any notions of gross
changes in the fluctuations of the magnetic field intensity
near the sea subsurface. An ice island is unique as a stable
platform from which to make magnetic measurements on the deep
ocean surface and at depth at the same time. The difficulties
in making such simultaneous measurements on other platforms in
the open ocean are known to persons who have attempted to make
measurements of this kind. Two ice islands were occupied by
the Arctic Geophysics group. Because of its larger base in¬
stallation T-3 (Fletcher1 s Ice Island) was chosen for these
measurements*

— 1

-2-

II. FIELD OPERATION

The magnetometer used was a modification of the Varian
Associates Modular V-4931 Proton Precession Station Magneto¬
meter.

This basic magnetometer is nearly identical to many other
proton precession magnetometers in common use. A hydrogen-rich
sample in the sensor is polarized by applying a relatively strong
magnetic field by current through turns of a solenoid surrounding
it. After polarization, the proton procession signal is induced
through these same turns. The frequency of the signal is deter¬
mined by allowing a predetermined number of signal cycles to
gate the output of a clock pulse generator of known repetition
rate. The number of clock pulses that get through the gate is
recorded on a digital counter. This reading, then, divided by
the number of preset cycles is the period of the signal. The
counter records the five low order digits of the output of the
pulse generator. The last two decimal digits are converted to
an analogue voltage and recorded on a strip chart recorder.

After the reading has been recorded the sample is again polar¬
ized, then the signal read, etc., in a cyclic fashion. The
precession frequency is directly related to the total geomag¬
netic field strength through an accurately known constant.

The instrument used in the present work had the follow¬
ing special features:

1) There were two sensors. Each sensor was polarized, then
read, then idle while the other sensor was polarized, and read.
For each sensor the polarize time was one second, the readout
time one second and the idle time two seconds giving an overall

-3-

cycle time of four seconds, A dual pen recorder was used with
each pen being activated with its sensor. The instrument op¬
eration is shown in the block diagram of Figure 1.

2) The clock pulse generator of the counter was changed from
100 kcps, as normally supplied, to a 200 kcps clock to get more
sensitivity (± 0.7 gammas). The full scale span of the strip
chart recorder was 70 gammas. On occasion the sensitivity was
degenerated because of noise but for the major part of the re¬
cording period a sensitivity of 0,7 gammas was obtained.

3) A kerosene-heptane mixture was used as a sample so that
the sample would not freeze at the subzero temperatures en¬
countered.

4) The surface sensor was attached by a 100 foot nonmagnetic
cable containing a type 310 stainless steel braid stress member.
The lower sensor was attached by a 10^0 foot cable. The fifty
feet nearest the head contained type 310 stainless braid and
should introduce no more than two gammas constant error in the
magnetic readings. The remaining part of the cable utilized
type 304 stainless which is more magnetic but did not intro¬
duce errors in the readings.

5) The housing of the lower sensor had pressure equalizing
diaphrams to reduce the detrimental effects of the hydrostatic
pressure at the operating depth.

The physical arrangement of the special field installa¬
tion is shown in the sketch of Figure 2 and in Plate 1, The
facilities were established at a location near the island edge
and isolated from the main camp. Except for one occasion when
ice rafted onto the island, this location proved satisfactory.

PLATE 1

PHOTOGRAPH OP UPPER SENSOR (IN THE DISTANCE)
AND THE HOLE THROUGH THE ICE FOR ENTRY
OP THE LOWER SENSOR (IN PORE GROUND)

22 E-*

r-

n£>

XA

-=J-

rr\

00

SECONDS

At that time prompt removal of the sensors prevented any damage.
The lower sensor was placed by a pulley over a hole through thin
ice. The hole was kept from refreezing by placing an electrical
heater wire in it. The major impedment to the operation was a
twenty day failure of the camp*s electrical generators when
spare parts could not be readily obtained. There were other
logistic and electronic troubles which are inherent in an oper¬
ation of this kind.

The installation began operation l4 October 19&2. Ex¬
cept for the twenty day period mentioned above it was possible
to record fairly regularly on the paper strip chart record.

Random and selected times were recorded on magnetic tape by the
use of retransmitting slidewires attached to the pen recorder.

On 26 November 19&2, recording with two s ensors had to be
temporarily abandoned due to equipment failure. Records and
tape recorder were returned to Lamont at that time. Electronic
parts have since been sent to T-3 and the two sensor, strip
chart recordings have been resumed. This report covers the
period from ll| October to 26 November 1962.

Fixes (weather permitting) and soundings were taken
once a day. The movement of the island between fixes is some¬
what questionable, but the fixes with straight line interpolations
between them is shown in Figures 3A and 3®*

III. GENERAL DESCRIPTION OF RESULTS

Although the ice land was north of the auroral zone it

o

ro

00

Flfr. 3B - POSITIONS OF T-3 31 OCT. - 20 JAN., 1962

5-

was located in a geographic area where intense time variations
of geomagnetic field are common* There were aurorae visible and
magnetic fluctuations accompanied them. Figure 4 shows the
magnetometer traces during one such interval. There were other
periods of several hours when the magnetometer traces were near¬
ly straight.

Hall (1962) has calculated signal frequency variations
that are to be expected if the sensing head of a proton preces¬
sion magnetometer is in rotation about an arbitary axis. On
several occasions the lower sensing head showed a rapid cyclic
change in signal frequency that was not evident on the upper.

See Figures 5> and 6. Calculations showed that these variations
could be accounted for by a slow rotation of the lower head
about a vertical axis. Subsequently, an intentional rotation
of the upper head about a vertical axis gave a similar result.

A single fin was attached to the lower head in an attempt to
prevent motion of this type, but the resulting drag raised the
head to such an extent that it was not deemed wise to use the
fin. Rotation of the lower head occurred on such infrequent
occasions that it did not interfere with general observations.

Appendix I shows how the effect of drag, caused by dif¬
ferential motion between the ice and sea water, can be estimated.
In a typical case the lower head may be only 890 feet below the
surface when 1000 feet of cable is in the water.

-6-

IV. TIME VARIATIONS

The instrument recorded the total intensity of the mag¬
netic field. The total field strength, P, is related to the
horizontal field strength, H, and the vertical field strength,
Z, by:

and small changes in these quantities are interrelated by the
equation:

F + AF = [(H + AH)2 + (Z + AZ)2]^

so that to a good approximation:

AF = (-pr)AH + (p-)AZ

= (cos D)AH + (sin D)AZ

where D is the dip. At the recording site the dip was approx¬
imately 87 degrees so that:

AF = (.052) AH + (.998) AZ .

It is clear that it would take a very large time fluctuation in
the horizontal intensity. Ah, to alter the total field intensity
to a measurable extent. It was the vertical component of time
fluctuations that was recorded.

The calculation of attenuations and phase shifts expected

-7-

in the vertical component with depth must embody the geometry of
the source and the curvature of earth structures. Such a devel¬
opment is beyond the scope of the present report, although it
is understood that such a development is in progress elsewhere
(A. T. Price, personal communication).

The assumption of plane electromagnetic waves normally
incident on the sea surface is unsatisfactory because such a
wave is not permitted, by Maxwell* s equations, to have a verti¬
cal time varying component (for example see Panofsky and Phillips,
1955)* The assumption of a non-normal plane electromagnetic
wave or the assumption of a hydromagnetic wave would introduce
an unwarrented degree of freedom to the calculation.

Two relatively short time intervals (approximately thirty
minutes each) of the records were subjected to auto and cross
power spectral analyses. The first section of record analyzed
is shown in Figure 7 (between the arrows) and its spectra in
Figures 8 and 9* The second section is shown in Figure 10 and
its spectra in Figures 11 and 12. Figures 9 and 12 include the
ratio of the power densities as a function of frequency.

Although an attempt was made to digitize the records from
the analogue magnetic tape recordings the similarity of the two
records and the limited dynamic range of the tape recorder pro¬
hibited an adequate digitization. Accordingly, the sections of
record were scaled by hand for a 1.25 second digitization inter¬
val. As Figures 8 and 11 show, the rapid decrease in spectral
amplitudes with frequency causes digitization noise to become
important for the shorter period activity. Those figures show
the part of the spectrum that can be considered free of such
noise, very conservatively estimated. Since each section of the

FIGURE 8

FREQ (CPS)

o vi asvai

O O 9

o o

vO o'* O

H

O

PERIOD (SEC)

iooo r

100

2

jo

CM

10

Q*1 —
0

POWER

SPECTRA

POWER
SPECTRA
LOWER

DIGITIZATION

NOISE

POSSIBLE

FIGURE 11
SPECTRA

1509.4 - 1539.4 z

26 OCT. 1962

80#

CONFIDENCE

LIMITS

CROSS

SPECTRA

.01

FREQ (CPS)

1

FIGURE 12

t

1

ovi asvHd

*

— i

o

o

o

O

ro

o

1A

T ^

N

T

PERIOD (SEC)

-8-

chart was digitized twice (for upper -and lower heads) there was
some possibility of a relative time displacement in the digitized
data* This would be manifest as an error in the phase spectra,,

In Figures 9 and- 12 the vertical lines with bars indicate the
phase error that would be so introduced, again very conservative¬
ly estimated*

The spectral determinations were made on an electronic
digital computer utilizing the general procedures of the auto¬
covariance method (Blackman and Tukey, 1958)® The individual
steps of the procedure were:

a) convert digital units to gammas

b) remove mean value; remove linear trend by using average
of first third and average of last third of data

c) filter to produce a nearly white spectra and to eliminate
alaising by periods of less than ten seconds

d) compute lagged products; series lagged 10/2 (20 degrees of
freedom)

e) compute power spectra, coherency, phase

f) '’ham’1 (smooth) the spectra

g) remove effect of filter

Between the periods of 70 and !|.00 seconds the spectra indicate:
a) for the longer of these periods the amplitudes of the lower
head are attenuated* While the limitations of the instru¬
ment and analysis do not give a clear indication of the
amount of attenuation as a function of the wave period there
is evidence for reduced attenuation or possible enhancement
of the lower head intensity for periods shorter than about
90 seconds

b) an increase in phase shift with frequency*

V. SPATIAL VARIATIONS

It is of some interest to determine the magnitude of the
gradient of the total magnetic intensity due to the earth1 s
main field* The magnitude of the geomagnetic field intensity
can be obtained from the expression for the magnetic potential*
Including only first order terms the expression for the magnetic
potential is:

V = (£r2 [g® sin A + (g1, cos B + h^sin B) cos a]

r is the distance from the center of the earth to the point of
observation, R the radius of the earth, A the latitude, B the
longitude, and g^ , g^ , and are constants. Since the

constants Q* and h* are no more than 20% of Q1 and since
the experiment was made at a high latitude (therefore QOS ^
small) only the first term in this expression need be retained*
This is equivalent to assuming that the earth* s magnetic field
is due to a dipole. For this case we use the common expression
for the total geomagnetic field intensity as a function of
radial distance and latitude

F = F0 (y)3 s/a cos 2 A + sin 2 A '

wl\ere F^) is the equatorial surface field value.

At the surface of the field installation r = R and

F = F,

F0 s/a cos2 A + sin2 A '

-10

and at any point below the surface of the installation

At a depth d below the surface r = R - d and

F * <l£-/ F. « " +

The difference between the lower and upper field strengths is:

AF = F - F, = F,(^)

Taking d = 1000 ft, R = 2.1 X 107 ft and Fx = 57,500 gammas
Af = 8 #3 gammas

In addition to the earth’s main field there are magnetic
gradients caused by geologic bodies at or beneath the ocean
floor. These bodies may have a remanent or induced magnetization.
The amount of remanent magnetization present is determined by
the structure’s magnetic history and cannot be calculated. How¬
ever, the gradient of the total field anomaly due to induced
magnetization can be calculated for an assumed body geometry.

This was done on an electronic digital computer for several two
dimensional structures following the method of Heirtzler, et al.
(1962). The results of one calculation (a fault with upper
surface on the ocean bottom) are shown in Figure 13* The mag¬
netic susceptibility of 0.006 used in this calculation yielded
an anomaly gradient of the type observed on 18 November 19&2
(see Figure 15)« This value of susceptibility is approximately
the same as that required by Hunkins, et al. (1962) in accounting

FIG. 14- MAGNETIC GRADIENT, DEPTH
AND GRAVITY ANOMALY PROFILES

-11-

for an anomaly over the Chukchi Cap,

The magnetic effects of geological bodies dominated the
effects of the main field as far as the magnetic gradients are
concerned. There were a number of occasions where the field was
of greater magnitude on the surface than at depth. If the earth’s
main field alone had been operative, the lower head would always
have given a higher field intensity. Figure l4 shows the mag¬
netic gradient, depth and gravity anomalies observed over a
sample period of five days. Due to the irregular track of the
island and shortage of good determinations of position, it was
not possible to produce reasonable contour plots of the data.
Figure 15> shows one of the more successful attempts to contour
depths and magnetic gradient.

VI. CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORE

With the instrument and recording technique employed it
was possible to measure the changes in vertical gradient as the
ice island drifted over geological bodies and to get general
estimates of attenuation and phase shifts of the vertical com¬
ponent of time variations between 70 and 4^0 seconds period.

There may be anomalous attenuations near 90 seconds period. How¬
ever, no gross attenuations were found.

For the future study of time variations beneath the sea
surface the sensitivity of the instrument needs to be improved
by (a) increasing the repetation rate of the clock pulse gener¬
ator and (b) by fixing the instrument in place on the bottom so
that no rotational effects are operative. At the present time
a self-contained bottom instrument with digital acoustic tele¬
metry is under construction at Lamont as an in-house effort.

166° W

30' N

15'

- DRIFT TRACK

- DEPTH ( M)

- GEOMAGNETIC GRADIENT (y/IOOO FT.)

FIG. 15- CONTOURS OF DEPTH AND
MAGNETIC GRADIENT, 9-23 NOVEMBER

12

Tiriis instrument uses a one megacycle clock pulse generator and
will have an accuracy of ± 0.1 gamma if the counter is recorded
digitally. With this increased sensitivity and with the increas¬
ed depth of the lower head attenuations and phase shifts will
be more definitely known and the spectrum can be examined to
somewhat shorter periods.

In high latitudes it will be advantageous to operate this
new instrument, part-time, with a magnetic bias field to cancel
part of the vertical d.c. field component. By this means the
horizontal component of time variations will play a more dominant
role and the two components of time variations can be studied
independently. The entire vertical field cannot be eliminated
since the resulting field would be too low to measure with this
type instrument.

At lower latitudes, however, one could cancel the entire
vertical component and study the horizontal time fluctuations
exclusively.

-13-

APPENDIX I

Configuration of Cable
Due to Drag of Water

Calculations were made to find the configuration of the
cable and sensor under conditions of uniform ice drift over an
ocean without other currents. Current measurements from pre¬
vious ice stations have shown that most of the change in relative
velocity between the ice and water occurs in a fairly thin
boundary layer just beneath the ice. The assumption of ice
moving over a motionless ocean is, therefore, valid as a first
approximation.

The problem was solved according to the technique out¬
lined by L, Pode in Report 687 of The David Taylor Model Basin,
’’Tables for Computing; the Equilibrium Configuration of a Flex¬

ible Cable in a Uniform Stream”. Pode tabulates certain ’’cable
functions” which are the numerical solutions of the differential
equation of the cable hanging in equilibrium under the influence
of a uniform current. It is assumed that the hydrodynamic force
which acts on an element of cable depends only on the angle that
the element makes with the stream and is not affected by such
matters as the curvature of the cable or the flow at neighboring
elements. This assumption is considered to be valid in this
application.

The sensor was considered to be a cylinder 20” long and
5” in diameter weighing 36 lbs. in water. The cable was con¬
sidered to be a cylinder with weight of 0.193 lbs/ft. For a
current velocity of l/4 ft/sec., hydrodynamic drag on the sensor

-i4-

and cable were calculated from the formula.

Drag = CD kP V2 where CD = drag coefficient

2 A = cross-sectional area

P - water density
V = current velocity

For the sensor, drag was 1.18 lbs. and for the cable it
was 0.122 lbs/ft.

Entry into Pode*s tables gave the depth of the sensor as
89O ft. and the horizontal displacement of the sensor as I4.I4.O ft,
for a cable length of 1000 ft. The accompanying diagram illus¬
trates the configuration. The diagram is not to scale.

/

-15-

acknowledgments

Dr. Kenneth Hunkins, head of Arctic Geophysics Department
of Lamont Geological Observatory, devoted a considerable amount
of time to the management of this project and to resolving the
technical problems that arose. Mr. James F. Cottone, of the Geo¬
magnetism Department, was responsible for seeing that all aspects
of the instrumentation system were technically sound and met
specifications. He installed and operated the instrument during
the three month field trip required to obtain the data. Mr*
Arthur Jokela materially assisted in the operation of the instru¬
ment, obtained other geophysical data mentioned in this report,
and analyzed the results to determine the effects of geological
structures. Mr. M. J. Davidson provided computer programs for
the statistical analysis.

The Arctic Research Laboratory at Barrow, Alaska, provided
the logistic support north of Alaska.

The project received financial support from the Naval
Ordnance Laboratory, Silver Spring, Maryland, through the Office
of Naval Research, Contract Nonr 266(82)*

Reproduction of this document in whole, or in part, is
permitted for any purpose of the United States Government.

REFERENCES

Blackman, R.B., and J.W. Tukey, 1958; The Measurement of
Power Spectra: Dover Publications, Inc., New York*

Hall, S*H*, 1982; The Modulation of a Proton Magnetometer

Signal due to Rotation: Geophysical Journal of R.A,S,,

Vol. 7, No. 1, pp. 131-141.

Heirtzler, J*R., G. Peter, M. Taiwan!, and E.G. Zurflueh,

1962; Magnetic Anomalies Caused by Two-Dimensional
Structure: Their Computation by Digital Computers and

Their Interpretation: Tech. Rpt. No. 6, Lamont Geo¬
logical Observatory.

Hunkins, K., T* Herron, H. Kutschale, and G. Peter, 19&2;

Geophysical Studies of the Chukchi Cap: Jour. Geophys.
Res., Vol. 67, No. 1, pp. 235-248.

Panof sky, W.K.K., and Melba Phillips, 1955; Classical Electricity
and Magnetism: Addison-Wesley, Reading, Massachusetts.